Second Case, Two Problems: GLP-1, Hypokalemia, and an Uncertain 'Allergy'
Mr. R, 68, is booked as the second case at 0930 for outpatient screening colonoscopy under propofol sedation. At 0800 preop review you find: type 2 diabetes with HbA1c 8.9%; BMI 34; weekly semaglutide 1 mg, last dose 6 days ago; a charted 'latex allergy - rash' after a childhood dental visit; and a basic metabolic panel from 5 days ago showing potassium 2.9 mEq/L and creatinine 1.1. He tells you he ate a light breakfast at 0700 'because I was hungry.' His wife adds that he once had a 'bad reaction' to numbing gel at the dentist. Vitals: HR 92, BP 148/88, SpO2 97% on room air. The charge nurse wants you to start the IV and send for the patient so the list stays on schedule.
1. Which preoperative findings must be escalated before this case starts, and in what order of urgency?
Read the worked answer
Escalate to the anesthesia provider and the endoscopist together, before an IV is started.
- Potassium 2.9 mEq/L (5 days old). This is the most actionable item. Hypokalemia lowers the arrhythmia threshold under sedation and can potentiate QT prolongation. It is also old - if he has been vomiting from delayed gastric emptying, it may be worse now. Ask for a repeat draw and an ECG, and let anesthesia decide whether to replete and proceed or reschedule.
- NPO violation. Breakfast at 0700 for a 0930 sedation is roughly 2.5 hours - far short of the 2-hour clear/6-hour light-meal standard. Elective sedation should be delayed, not 'worked in.'
- GLP-1 agonist. Weekly semaglutide was last taken 6 days ago, which does not meet the recommended one-week hold for weekly agents; combined with an elevated HbA1c (suggesting delayed gastric emptying), treat him as a full stomach rather than assuming he is empty.
- Allergy history - clarify before any drug is drawn up.
I would tell the charge nurse plainly: the schedule can move, but I am not starting this case without an anesthesia assessment.
2. How would you sort out the two reported 'allergies,' and what would you document?
Read the worked answer
Treat them as two separate, unverified reports and do not let either become a permanent label without evidence.
- Latex 'rash' after a dental visit: a rash is not automatically IgE-mediated, but the cost of a latex-safe environment is low and the consequence of being wrong is anaphylaxis. Reasonable plan: keep the latex-free cart and gloves, run him as the first case of the day, and document the report verbatim with the reaction described.
- 'Bad reaction' to numbing gel: find out which agent (benzocaine is an ester, lidocaine an amide) and what actually happened - a true reaction versus a vasovagal episode after injection. Do not chart 'allergic to all caines'; amide-ester cross-reactivity is low, and an overbroad label limits future care.
- Record the patient's own words, the date, the agent if known, and the treatment required, then flag both to anesthesia and to the endoscopist before the propofol is drawn up. Recommend allergy clarification or referral as an outpatient issue.
3. If anesthesia decides to proceed after discussion, what aspiration precautions would you advocate, and why those specifically?
Read the worked answer
The combination of a GLP-1 agonist, poor glycemic control, and a recent meal puts him in a high-risk category that a 'standard' propofol sedation does not address.
I would advocate for:
- A full-stomach approach: rapid-sequence induction with endotracheal intubation rather than deep sedation, or at minimum a point-of-care gastric ultrasound to guide the decision.
- Working suction at the head of the bed and a second suction available, plus head-up positioning for induction and extubation.
- Continuous capnography and end-tidal CO2 throughout, which is the earliest reliable signal of an obstructed or failing airway.
- A setting with rescue capability. If the endoscopy suite cannot support an unplanned intubation, that is itself a reason to move the case.
My reasoning: the risk here is not the sedation drug, it is stomach contents in a patient whose emptying may be delayed. Everything above is aimed at that single mechanism rather than at the monitor numbers.